How empirical uncertainties influence the stability of climate coalitions

International climate agreements are negotiated in the face of uncertainties concerning the costs and benefits of abatement and in the presence of incentives for free-riding. Numerical climate coalition models provide estimates of the challenges affecting cooperation, but often resort to assuming certainty with respect to the values of model parameters. We study the impact of uncertainty on the stability of coalitions in the Model of International Climate Agreements using the technique of Monte Carlo analysis. We extend the existing literature by (1) calibrating parametric uncertainty about damages and abatement costs to estimates from meta-studies and by (2) explicitly considering uncertainty in the curvature of the damage function. We find that stability is more sensitive to uncertainty in damages than in abatement costs and most sensitive to uncertainty about the regional distribution of damages. Our calculations suggest that heterogeneity can increase stability of coalitions; however, this depends on the availability of transfers.